Insulin resistance occurs in a variety of pathological states including obesity, non-insulin dependent diabetes mellitus (NIDDM) and hypertension. Decreased glucose effectiveness has been reported in NIDDM. Since insulin resistance and impaired glucose effectiveness may play a role in the pathogenesis of NIDDM, quantitative assessment of these properties is of interest. Recently, a minimal model approach involving mathematical modeling and computer simulation has been used to estimate both insulin sensitivity and glucose effectiveness from the results of a single frequently samples intravenous glucose tolerance test (FSIVGTT). The equations of the minimal model describe changes in plasma glucose concentration as functions of insulin and glucose concentrations. A computer program identifies model parameters that generate a best fit to insulin and glucose data obtained during the FSIVGTT. Thus, the minimal model is able to estimate the relative contributions of insulin and glucose to glucose tolerance. We have used the minimal model to generate specific predictions of both insulin-dependent and insulin-independent glucose metabolism under a variety of conditions. We performed experiments in subjects with insulin dependent diabetes mellitus and in normal subjects to test these predictions. By comparing model predictions with experimental results, we were able to demonstrate that the minimal model underestimates the contributions of insulin and overestimates the contribution of glucose to glucose metabolism. Studies are presently underway to confirm these results and to understand the origin of the discordance between minimal model predictions and experimental results. The data that we collect may allow us to formulate a more accurate and useful mathematical model.